The overcurrent protection is pretty quick, it might have triggered on inrush surge. The USB chip on my scope (USB2517, the proprietary probe interface is just USB + 2 pins) supported the following timings:
5:4 OC_TIMER Over-Current Timer: Over-Current Timer delay.
00=0.1ms
01=4ms
10=8ms
11=16ms
You can see from my measurements (17.8ms cutout) that it was programmed to the longest delay, not shortest.